ABSTRACTS OF COMPLETED DOCTORAL PROJECTS

Frames of Reference and Direct Manipulation Based Navigation -Moving in Virtual Architecture Space.
A research project undertaken by A. Friedman (2005).
Although highly advanced in visual realism, current virtual reality systems used by designers do not provide the user with the ability to interact effortlessly and move in space as desired as one does in real space. As a result, these systems offer limited help to designers when they want to test or communicate the spatial quality of their projects. The aim of the research was to develop a design tool for navigation in virtual environments that can offer a sense of ‘immersion’ and vividness similar to that experienced when one moves in real space.
The research examined a number of typical cases of CAD-VR navigation systems and analyzed their performance. All programs use direct manipulation: a virtual representation of reality is created, which can be manipulated by the user through physical actions like pointing, clicking, dragging, and sliding. The analysis showed that the lack realism of these systems means that they do not offer a ‘sense of reality’ because of the user inability to interact easily with the computer to navigate among represented objects. The user cannot: 1. Plan a course from a given location to a desired one. 2. Shift the direction of their gaze and focus attention to objects as they move across a path 3. Move around an object keeping truck of this change in relation to the surrounding objects. 4. Turn an object in front of the viewer in order to examine it. This lack of ‘sense of reality’ cannot be improved by adding attributes that are more realistic - details, shadows, and reflections. Departing from pioneering rigorous studies developed in environmental design by Kevin Lynch and his followers about ‘cognitive mapping’, and drawing by recent research in cognitive science on spatial thinking, the study identified these aspects which acknowledged the cognitive structures and processes through which people perceive their environment as they move through it. In contrast to Lynch’s approach concerned with visual quality of urban environments focusing on visual urban cues for recognition and orientation within a city, the present research, related to movement through the built environment, concentrated on the ‘frames of reference’ people use to plan their path among objects, shift their attention to them, move around them, and turn them. The frames of reference used are 1. allocentric, 2. egocentric, 3. relative, 4. intrinsic, and 5. absolute. Following the criteria of realism and immersion in exploring through movement the virtual world, the system uses an agent that allows the user to direct navigation and view of objects. It permits for both agent-based and object-based navigation. The user can refer to the movement of the agent as a base for movement or to the object as a reference point. To enhance the feeling of vividness the user’s input to the system requesting change of viewing position, direction of view, object of view, or a path, or a view, is expressed in terms of natural language while the output remains visual. In other words the user talks to the agent and sees on the screen what the agent views in the virtual world. The user-centered navigation tool produces “on the fly” navigation most desirable for design pro fessional applications as opposed to a tailored presentation. It can be applied in urban environ ments as well as in architectural interiors both using the same types of axes and frames of reference. It is targeted to support testing the quality of designed environments, both interior and exterior, by individual designers but it can be most effective in architectural presentations and debates where the architect communicates with various parties, while examining the various as pects of the three-dimensional project.

A Tool for the Design of Facilities for the Sustainable Production of Knowledge.
A research project undertaken by Jun Wu (2005).
Every epoch has its leading type of building. Today, given the leading role new knowledge plays in the world economy, innovation-producing facilities such as universities, research centers and scientific parks emerge as the leading building type of the 21st century. The research focuses on developing a design tool to facilitate the design of such facilities whose primary characteristics is long-term, sustainable creation of new knowledge.
The research draws from current research on the subject of scientific, technological, and cultural innovation which stresses the importance of highly diversity clusters of face-to-face interacting people as a necessary condition enhancing sustainable innovation. The research investigates the necessary physical building conditions that permit the flow of information in such clusters. It examines the topological and time-distance constraints of the physical environment that affect the potential information flow inside such clusters and among clusters, and tries to identify properties of the topological organizations as they influence possible physical interaction.

There are three aspects where the design tool is pioneering:
1) Development of models to represent and control face-to-face communication between groups in the built environment, with the aim to enhance group diversity of interaction.
2) Development of advanced multi-goal adaptive system applied to the organization of building com plexes such as universities and research centers.
3) Transfer of concepts and techniques used in models of economic innovation and environmental sustainability, to the area of social quality of the environment and face-to-face interaction.
The research is multi-disciplinary. It employs methods developed in regional science, sociology of scientific groups and innovation, architectural theory, and design methodology. Studies in the field of biology are of particular interest because they deal with relations between environmental conditions, adaptation and diversity having structural similar ties to the problem of knowledge production. The final goal of the tool is to identify physical patterns, spatial arrangements that we define as ‘Archigraphs’. It provides design guidelines to construct the topological network of facilities to be applied at the early stages of the conception of design process.

Emergent Places for Urban Groups without a Place - Representation, Explanation and Prescription.
A research project undertaken by S. Vyzoviti (2005).
Urban Groups: Representation; Explanation; Prescriptions In the past few years we have been observing increasing appreciation of the presence of groups that have not been accounted for in the design of urban spaces, which we call non-place urban groups. The goal of the dissertation is to investigate: How informal urban groups, constituted on the basis of social and cultural characteristics that denote shared systems of norms, values, interests, beliefs and behaviors, appropriate space sponta neously. Ways of representing sequences of activities carried out by such groups as they form networks manifest in the development of public space ‘bottom up’, outside an institutionalized framework. Based on the results of the above investigations leading to a model, the dissertation develops design guidelines that specify environmental necessary conditions contributing to social interaction in public space, for specific targeted groups.The research method draws from empirical observation techniques to identify data concerning the
use of space.
The research is interdisciplinary, drawing knowledge from culture studies, urban sociology, social geography, environmental psychology, and architecture theory and design methodology. The dissertation benefits from prior theoretical work by: urban sociologists, geographers and scientists. The research employs two case studies. Focusing on the spontaneous appropriations of circumscribed urban spaces with certain continuity in time, two emergent urban groups are respectively studied in two urban situations. The purpose of the first case study, the Immigrants’ Place of Getting- together in Athens Downtown is to provide the empirical data upon which a behavioral and a normative model are constructed. The second case study, the Skateboarders’ appropriation of the Tilburg Conservatorium public grounds, is used as a test case for the application and evaluation of the models. Anchoring on empirical observation the dissertation produces two kinds of models of emergent places: a behavioral, descriptive – explanatory model and a normative, prescriptive - proscriptive model. The models relate to two scales of the urban environment: the urban regional pertaining to the selection of the location and the urban microlocational, pertaining to the organization of activities within the site of the location.
Besides systematic representation the intention in the behavioral model is to identify the minimum necessary environmental conditions contributing to the emergence of the specific social interaction spaces. The normative model, which is situated and constructed upon the basis of the behavioral one, comprises a set of design guidelines for the accommodation of ‘urban groups without a place’. The pragmatic value of the research in design and planning is the enhancement of social quality in public space, by understanding the presence of ever-emergent urban groups in the context of accelerated cultural change. The innovative character of research, besides comprising a fresh gaze to the problematic of the user, is that it provides a tool for accommodating environmental needs for not yet specified spontaneous groups using urban space.

CaseBook: Toward retrieval by lazy classification of architectural information objects
A research project undertaken by. Sinan Inanc. (2003). It involved the accessibility of the large amounts of information objects available from various repositories becoming a major problem in architecture due to the absence of suitable information retrieval (IR) environments. The heuristics of lazy classification is proposed in order to improve the usually poor retrieval performance caused by the limited level of correspondence among rigidly classified information objects and the difficult-to-predict, ever-changing perspectives of the users.
Most IR environments rely on rigid, manually created classification of information objects according to conventional library or hypermedia paradigms. Content-based IR approaches promise to address problems associated with manual processing, such as ambiguity, information loss or labor costs, by automatic processing of the information objects. The lazy classification heuristics is in principle similar to existing content-based clas sification techniques, except for the emphasis on deferred processing of the information objects. The main advantage of feature extraction during the IR interaction instead of pre-processing is the possibility to guide the classification processes according the preferences of the inquirer in order to increase the responsiveness of the IR environment. The major drawbacks are the complexity of the system development and the required computational resources during the IR interaction. In return, lazy IR environments offer the information suppliers the appealing perspective of the possibility to avoid some ungrateful classification-related tasks and the increased responsiveness creates the possibility of attracting and holding the attention of larger audiences. The successful implementation of lazy classification depends on the availability of the formalizable domain knowledge that is applicable for the development of automatic feature extraction algorithms, whereby the efficiency for on-demand processing and the adjustability for user guidance become minimal requirements. Additionally, the capability of influencing feature extraction during the IR interaction increases the demand for reflective user interfaces for navigating the repositories without drowning in the necessarily increased
complexity of information content and navigational tools.
An experimental IR environment, called architecture CaseBook, is implemented for the examination of the feasibility and exploration of the potential of the advocated heuristics of lazy classification in a typical archi tectural domain. CaseBook is an IR environment for residential unit floor plans that operates on the similarity of spatial characteristics. Manual processing is unfortunately not completely eliminated but human interven tion is used to create formalized diagrammatic surrogates of the floor plans that are suitable for automatic processing. The chosen level of detail of the formal diagrams in CaseBook is such that many visual and spatial characteristics of the original floor plans can be captured while the semantics used makes the auto matic extraction of various spatial features feasible. The capability of extracting a relatively large set of spatial features from diagrammatic representations of floor plans gives the powerful possibility to the usage of floor plan diagrams as the core internal representation for storage and querying. The applicability of lazy classification is studied through partially adjustable automatic extraction modules for spatial features. A set of processing parameters, here termed hinder values, can be altered through a simple interface in order to personalize the interpretation of the spatial organization of the floor plans. The performance problems that can be expected as a consequence of the on-demand classification are addressed through incremental processing and multi-level caching strategies. In order to examine the feasibility and potential of lazy classification heuristics in a realistic IR environment, the retrieval of floor plans is further based on similarity assessments by means of a weighted multi-dimensional vector representation and a query-by-sketch interface to ease the IR interaction.

Use and Adaptation of Precedents in Architectural Design
Karina Moraes Zarzar (2003) researched the problems of re-use of design precedents in the conception of new design solutions. The goal of this research project was to contribute to the construction of computational tools to facilitate this practice by developing a model that grasps significant characteristics of the design process as it employs precedents. The model was built drawing an analogy from the natural evolution. The intention was not to represent the processes that take place in the architects’ minds but rather their behavior as this is manifested in their design products.
The project drew from the multidisciplinary methodology of design tool development of the Design Knowledge Systems Research Center. It employed an analogy with Darwinian evolutionary theory in combination with recent theories of genetics and embryology. The criteria of usefulness in picturing the phenomenon in architecture determined the focus on particular aspects of the analogy. The research used three case studies from the architectural domain: J.J.P. Oud, to identify adequacy criteria for the model; Le Corbusier, to illustrate the components and conduct of the model under development; and Santiago Calatrava, to test the model. The research developed a pre-computational qualitative model that provided insights into the process of re-use by analogy of precedents in architectural practice.
Given the notorious history of misuse of biological models applied to other fields, special attention was paid to circumscribe the limits of the analogy and recognize the differences between design and evolutionary models, the most important being the process of selection, natural versus artificial. As “breeders”, designers recall from memory and/or from archives through “artificial selection” - this is not the case in natural selection. In natural evolution, mutations are “random”, and natural selection gives the direction. In the human design process, mutations and selection are mostly intentional.
The evolutionary and genetics analogy serve as heuristic devices to represent the mechanisms in the process of use and adaptation of design precedents and the elements of such precedents accumulated over the years that are adapted and recombined during the design process often leading to design innovations. The model employs the notion of “design feature”, a precedent component, as the most important unit of selection. Drawing from developmental genetics, and the idea of regulatory genes, each feature is derived from two interlinked kind of instructions where the “regulatory d-gene” deals with the configurational instructions and the “structural d-gene” deals with the technique and materials used. In the design model, just as in evolution, the notion of fitting environmental constraints in the generation of form is essential. Fitness relates to both internal and external constraints; it is multi-dimensional in a multi-criteria ecological environment.

Cognitive biases and errors in spatial design thinking: The case of design biases A research project undertaken by Philip Joo Hwa Bay, (2001). It investigated: 1) How cognitive biases (or illusions) may lead to errors in design thinking 2) Why architects use architectural precedents as heuristics despite such possible errors. 3) The mechanisms of a design tool that can overcome this type of errors and
improve accuracy in architectural thinking.
The research method applied was interdisciplinary. It employs knowledge from cognitive science, environmental engineering, and architectural theory. The case study approach is also used choosing the practice of Tropical Architecture as a subject of analysis. The investigation of architectural biases draws from work by A. Tversky and D. Kahneman. The use of heuristics of representativeness (based on similarity) and availability (based on ease of recall and imaginability) for judgement of probability can result in cognitive biases of illusions of validity and biases due to imaginability respectively. Incomplete information, limited time, and human mental resources make design thinking in practice difficult and impossible to solve. It is not possible to analyse all possible alternative solutions, multiple contingencies, and multiple conflicting demands, as doing so will lead to combinatorial explosion. One of the ways to cope with the difficult design problem is to use precedents as heuristic devices, as shortcuts in design thinking, and at the risk of errors. This is done with analogical, pre-parametric, and qualitative means of thinking, without quantitative calculations. These heuristics can be efficient and reasonably effective, but may not always be correct, because they can have associated cognitive biases that lead to errors. Several debiasing strategies are discussed, and one possibility is to introduce a rebuttal mechanism to refocus the designer’s thinking on the negative and opposite outcomes in his judgements in order to debias these illusions. This strategy was tested with an experiment. The results show that the introduction of a rebuttal mechanism can debias and improve design judgements substantially in environmental control. The tool developed has possible applications in design practice and education, and in particular, in the designing of sustainable environments.

A Cognitive Framework for an Urban Environment Design Tool
Predrag Sidjanin (2001) provided a computer-based tool for designers to improve the visual quality of the environment and resolve visual-spatial conflicts and disfunctionalities resulting from the increasing complexity of the human-made environment. The tool draws from recent developments in urban and landscape theory, cognitive science, and electronic information systems.

The design of buildings has become increasingly collaborative. To meet this challenge, architects and their partners have found it necessary to devote resources to coordinating their activities. The means of this coordi nation, however, have remained mainly intuitive and informal. theory. The research of John L. Heintz (1999), has led to the development of a tool, the Design Collaboration (DeCo) system to support collaborative design by combining a communication design network with a series of game conflict resolution devices. The research is grounded on an empirical case study carried out by Peter Donker. A key feature of the model is that design participants have individual goals and criteria for success that may be in conflict with the goals and criteria of others, and they are free of any overall authority. This situation creates the need for the development and application of tools to guide the informal and moral relationships between the actors, and assist them in their goal of coordinating their actions. The concept of the social contract, especially as developed by John Rawls, is applied here. Social contract theory provides an account of how to determine what sorts of arrangements for the distribution of justice the individuals within a society will find acceptable. The collaborative design team is compared to a society engaged in a particular practice: collaborative building design (CBD). A procedure is then developed to show what sorts of coordination techniques members of the team would find acceptable. Principles of planning theory, speech act theory, and Kantian ethics are used to develop a model of the conditions necessary for effective collaborative design. These conditions include good faith in promise-making. This, in turn, requires knowledge of the goals and partial plans of the other actors. DeCo has been developed to facilitate the coordination of CBD. The first component of the DeCo system is the Collaborative Design Project Network (CDPN), based on the CBD model. The CDPN permits actors in the design process to define their participation in the project as they see it, and then knits these partial plans together to show how the work must be coordinated. Individual actors may coordinate the partial plans of several projects to achieve improved resource balancing without fear that their internal business decisions will be subject to outside scrutiny. The CDPN makes use of two elements in the project network: tasks and infotems-specifiable items of infor mation that are generated by and required for tasks. This allows many types of task prerequisites to be reduced to structural features of the network. The second component of the DeCo system is a set of two game-theory tools to facilitate the resolution con flicts that may emerge in the formation of the project network: a simple two-person game for assigning tasks claimed by two actors, and an N-person scheduling game. These tools aid in reaching a feasible project plan and an acceptable schedule. The taskassignment game is a two-stage non-zero-sum game that assigns tasks based on the preferences indicated by the actors. The scheduling game is modeled on Rawls’s game theory procedure for determining acceptability. Actors take turns proposing schedules while the others accept or reject them. The requirement of unanimity assures the fairness of the resulting schedule. The actors arrive at a generally acceptable schedule by learning each other’s needs through observation of the proposals the other actors offer.
A series of thought experiments based on scenarios derived from the case study are used to evaluate the performance of the DeCo system. A Rawlsian acceptability game procedure is then used to demonstrate the acceptability of DeCo. The DeCo system is immediately applicable as a set of procedures for improved design collaboration. A brief outline is also given of how these tools might be instantiated in the form of a system of software agents. The project provides both new scientific insights into the informal structures and processes of collaborative design and an important prescriptive model to be applied in design practice.

A second project in the theme of Collaborative Design was carried out by Peter Donker (1999). It departed from the same observation that architects have to face the challenge of designing as a complex interactive, interdependent activity, a fact resulting from increasing technical specialization, and globalization and the introduction of new electronic media.Where formerly architects might complete a preliminary design before engineers and other specialists are consulted, they are now requested to involve specialists early in the design process. This leads to more participants in the design process and more meetings, which in turn result in more communication “channels” and “forums.” The amount of communication between participants is, therefore, increasing. If we add to this the trend of increased litigation in our society as a whole, it is not surprising that design team participants find it necessary to be more meticulous in documenting the communication between themselves and other involved parties. Not unrelated to the phenomenon of specialization is the increased distance across which the design team members are required to cooperate with one another. The introduction of new electronic media to facilitate communication promises to be of help for these geographically dispersed design teams. Until now, however, the use of new media has generated problems rather than facilitated the process.

The study has lead to the development of the tool Structuring Communication in the Architectural Forum For On-Line Design (SCAFFOLD). The study focused on coordinating work within an architectural design team. It captured the communication that one encounters within a design team, so that it could be used to retrieve a previous discussion. In this sense it modelled the archive of a firm involved in a design project. Recapitulating discussions quickly is of importance to the coordination and monitoring of the design team’s efforts, to design decision-making, and in liability issues. The result of this work is a message information system that can be used by a design team leader to keep an overview of what the design team members are communicating. For this structuring a model is developed. The model represents those aspects of design team communication to facilitate documenting of design teams’ communication. It structures this communication so that they can revisit this material and learn why decisions were made the way they were, what the consequences are, what alternatives existed, etc. Moreover it is intended to help the decision-makers in the design process see where the design team partici pants stand on the issues and who has made what commitments. Two case studies were con ducted: the Almelo public library and the SABU-huis redevelopment project. Both projects are designed by Mecanoo architects. The Almelo public library case was studied retrospectively; i.e. through consultation of company archives. The SABU-huis was observed through the presence of the researcher at the design team meetings.
The model includes 1) a reasoning layer 2) speech-act layer overlays the reasoning layer. The tool developed can be used for recording professional meetings in design practice. It can be used as a database that can be consulted by those in volved in the design process. It is also useful for the development of protocols of electronic exchanges between geographically dispersed design teams.

Building Community: Design in the Organizational Mind was the subject of the research by B. Joseph Press (1999). In the search for meaning, the architectural profession legitimately seeks culture to sanction its products. However, in business organizations, culture is complex and tacit- richer and deeper than any of its external manifestations, including architecture. To compensate for culture’s incoherence, the profession assumes facile access to culture through existing artifacts and spatial usage. I contend this response limits the profession’s ability to engage social complexity, imbue architecture with cultural relevancy, and establish competitive advantage. This dissertation aims to provide insights into architectural form and process in relation to organizational culture. Schön contends tacit frames revealed in design activity circumscribe organizational culture. Further, the institutional and cultural status of these frames requires engaging in a collaborative design process. ‘Appearances of form’ in design activity demonstrate the presence of frames and simultaneously enable speculating about their tacit nature. Similar to the construction of frames, the design of an evolving physical object reveals how prior knowledge is assembled to facilitate sensemaking. Design in a social setting- characterized by negotiation, conflict, and agreement- sparks the frame restructuring required to coordinate disparate agendas through organizational learning.

Designing within the ‘collective memory’ and supplemented by the theory of type, design can leverage its potential to enlighten and improve organizational culture. Beginning with what designers share, the practices of Louis Kahn demonstrate cultivating an ‘archi-type’- form containing both cultural and architectural knowledge. To imbue each with ‘good’ form, the architects collaboratively creating organizational space to direct architectural form and redirect cultural action.

By seeking shared understanding through form, architectural design stimulates organizational reflection, learning, and agreement. Implanting these virtues occurs by an architectural design process stimulating the emergence of culture though ‘bricolage’- the synthesis of current and future concerns with an omnipresent past to guide daily interaction. As form emerges, the architect encourages an organization to reassess the frames circumscribing its cultural activity. Heightening the appreciation and awareness of culture instills communal practices of cooperation, respect, and learning. To achieve such acumen and influence, however, requires ‘reframing’ our professional agenda to reinvigorate the cultural significance of architecture and the design process.

The interrelationship between the poetics of architecture, rhetoric, and music as expressed in the Renaissance garden design of Niccolo Il Tribolo was studied by Anna Beltrami. Prof. Paolo Carpeggiani is special advisor in this study.The interrelationship between the poetics of architecture, rhetoric, and music as expressed in the Renaissance garden design of Niccolo Il Tribolo was studied by Anna Beltrami. Prof. Paolo Carpeggiani is special advisor in this study.

The partially published writings of S. Stevin on architecture were studied by Heidi de Mare. The research explores the knowledge structures underlying these texts and drawings to reveal the role of belief systems in giving shape to Stevin’s theories. Collaboration with Prof. W. Frijhoff.

Rob van der Bijl (1998) developed an argumentation structure suitable for addressing the problem of buildiong security. Decision-making in the field of burglary prevention can be supported by knowledge systems. Originally the knowledge domain of these systems was represented as a set of rules. It is conjectured that the case-based paradigm offers an efficient means of structuring design decision systems that are usable in the practice of crime prevention. Precedents (cases), as efficient ‘chunks’ of knowledge, offer some major improvements in the practice of building as well as using support systems. The research examines the fol lowing topics: (1) burglary prevention, (2) rules versus precedents/cases, (3) decision-making in crime prevention, (4) potential users of support systems, and (5) burglary detective support. Finally, a sketch is given of a burglary prevention system.

Precedents and Design Thinking in an Age of Relativization, France 1650 and 1793 Denis Bilodeau (1996) examined the relation between design cases and general abstract principles in the systemization of architectural knowledge in the emergence of contemporary architectural theory. His research concentrates on the debates around the relative value and power of concrete precedents and abstract principles in the definition of rules, prototypes, models, and design methods in French architectural discourses between 1650 and 1800. He proposes a reevaluation of the ideological debate on the relation between antique and modern architecture in the given period from a cognitive point of view and in the context of the conceptual constraints of the time. He argues that the tension between a priori abstract principles and empirical experience data that characterized the problem of systemization of architectural knowledge today originates from the crisis that occurred in the seventeenth and eighteenth centuries as a result of a redefinition of the value of the classical tradition. While in the middle of the seventeenth century, classical precedent cases appeared as a source of authority in decision-making due to of their antiquity and the sanctity of the tradition, they were gradually redefined as a source of experience and objects of critical reasoning and evaluation over the eighteenth century. During this process, the way the use of precedent was prescribed in architectural treatises evolved from a principle of imitation applied to particular canonical objects to the idea of precedent as a source from which to extract principles and generic ideas to be used by analogy in new designs.

Artificial Intelligence for Automated Floor Plan Generation
Davood Chichian’s research (1996) applied the strategy of hierarchical problem solving and decomposition to the generation of large-scale and high-complexity architectural floor plans, such as those of hospitals or education buildings. Simple manageable design tasks were considered individually and in relation to corresponding, similarly derived, tasks. Algorithms were developed to conceive plans at the topological level while simultaneously accommodating geometrical constraints. The Galois Lattice was employed as a mathe matical means to predict performance with respect to people sharing the locations. The system first con structed a tree structure from bottom up: from the root node it moves towards the leaf nodes. At every step some details were added to the partial plan until there was no more information to be added.

A Dialogical Model for Participatory Design: a computational approach to group planning
Hoang-Ell Jeng (1995) developed a computational model of participatory design dialogue. In participa tory design, design concepts are generated collectively through discussion, dialogical interactions in which the interchange of normative and supporting factual descriptions builds a collective design discourse. The goal of this research is to develop a method for participatory design to support this collective, face-to-face design problem-solving, in order to increase the acceptability of the design product. Since the mid-1960s, there has been an important movement towards increasing the participation of citizens in determining their built environment. At first, the movement was associated with social-political ideologies and rhetoric. By the end of the 1970s, participatory design had become an accepted component of professional practice. The objectives of the movement became more pragmatic and more modest, focusing on ex changing practical information, resolving conflicts, and supplementing design. Today, participatory design is in a new phase. Traditional participatory design methods are seen as insufficient to fulfill an increasing demand for dialogue. The point of departure of the study is the assumption that new information technologies can satisfy this demand.

The study focuses on the very early stages of the design process: the generation of design guidelines. It investigates the process of group planning and develops a computational model for the further realization of computer-based information systems to support that process. To develop this model, the study draws knowledge from cognitive science, argumentation theory, decision theory, and artificial intelligence. The study uses a case study as a heuristic device. The project selected is the participatory design of Ho- Chu-Wei Park in Taipei, Taiwan, in which Jeng was involved. The case serves to guide the development of the model. A design project in Amstelveen, the Netherlands, designed with the participation of people representing various interests, was selected as a second case to test the produced design-supporting method. The method includes a group-reasoning model, a dialogical system, and a framework for participation-based design guidelines. The group-reasoning model formulates the process of knowledge acquisition, the learning and sharing of belief systems, the generation of design alternatives and design evaluations. The dialogical system provides a clear description of how the information should be processed, what aspects should be paid attention to, what results can be anticipated, and when and how to control the process. The framework for participation-based design guidelines guides and structures the design processes. It facilitates a reconstruction of the implicit cognitive structure that underlies dialogue.

Rationalization, Standardization and Control in Institutionalized Design Reasoning
Peter Scriver (1994) studied the thinking processes and structures that helped produce the distinctive archi tecture and settlement planning of British India in the 19th century. The work offers insights on the rationalization and institutionalization of design thinking and the role of socially constructed conceptual systems in such processes in a multicultural context. The project was undertaken with the collaboration of Prof. Anthony D. King. The study examined the efforts of the Public Works Department of British India to ration alize the conception and execution of the wide range of civil and military buildings required by that colonial administration. This bureaucratically organized and regulated design system enabled a small number of engi neers and subordinate technicians, widely distributed geographically, to generate a large volume of formally consistent and serviceable architectural solutions with relative efficiency. A cognitive historical analysis of the relationships between this utilitarian architecture and the political and cultural goals of the colonial regime was pursued through a multifaceted examination of the processes and patterns of reasoning of the design engineers concerned. This was reconstructed through extensive research in colonial government archives in India and England. In the cognitive analysis of the case, the systematic design methodology of the public works institution was interpreted as a prototypical rule-based belief system, controlling professional action. These included the role of architectural design theory and methodology, such as canons of spatial composition, design routines, and architectural program stereotypes, as conceptual frameworks, in structuring and delimiting the design reasoning process; the codification of precedent knowledge in the form of design regulations and standard plans; the role of belief systems in controlling the way expert knowledge may be organized and employed institutionally.

Focusing on the case of the Unité d’Habitation of Le Corbusier, Prof. A. Tzonis has studied the use of analogy in architectural problem-solving employing precedents. He had the assistance of E. Offermans, archival research, Mark Cohen, computational representation of the process, D. Giannisis hyper-media techniques to apply the process in practice.

A Framework for Comparing and Controlling Number-based Design Reasoning Systems
Li Yu (1994) examined the role of number in the representation of building design knowledge. Two culturally distinct design number systems were compared: a traditional Chinese system employed in the classic Chinese house-builder’s manual, Zhai Pu Zhi Yoa (1741); and Le Modulor (1954) a modern system developed by Le Corbusier for universal use by contemporary architects. The comparative study explores the commensurability of these two systems of design-thinking, with particular regard to the association of numbers with formal and performance criteria in cognitive categories. The investigation focused on the means by which each culture categorizes design objects. In particular, the investigation explored three major issues: (1) design argumentation structure as found in two culturally distinct architectural measurement systems, (2) the definition of the possible obstacles to the communication between the two systems, and the means by which these obstacles might be overcome; and (3) the design of the System of Identifying Patterns of Design Arguments (SIPDA) for transforming design arguments from text into a specified form of abstract description, which captures the characteristics of a particular design reasoning process in cross-cultural contexts.
The results of the study develop a general computational system of design representation capable of handling cross-cultural situations. The system can be extended in design of object oriented programming applications. The study was carried out in collaboration with Prof. K. Ruitenbeek.

A Computational Representation of the Spatial Organization of Residential Buildings
Ir. Marc van Leusen (1994) has developed an automated design support system for the design of residential buildings. The system provides information interrelating spatial form types with operational and performance requirements. The project was undertaken in collaboration with Prof. K. Rijnboutt and Dr. Phillip Steadman. At the early stages of the design process, fundamental decisions are made with respect to the spatial organization of a project concerning overall shape, size, and internal organization. Such preparametric decisions, particularly concerning the building’s spatial organization, may have serious consequences for various aspects of the design product’s performance. The study aimed at supporting decisions at the early stages of the design process by developing a typology of pre-parametric possible design solutions. In addition to a review of existing work in the area of housing typology, the study develops a type-representation of basic arrangements of dwellings, which retains only the most general characteristics of spatial organization, and a demonstration of the potential of this representation in relation to design.

Meaning of the Site
The development of a system for classifying and analyzing building sites, taking into consideration multiple points of view and cultural belief systems, was the topic of the doctoral research of Xiaodong Li (1993). The research analyzed two systems of site analysis in a comparative manner: a traditional Asiatic system and a contemporary western system. Chinese FengShui theory, as described in the 15th century text Xiao Puo Ji, is a unique system which has been practiced in China and other Asian countries since the fourth century B.C. This theory was chosen for two reasons: FengShui provides both written and pictorial documentation on various kinds of sites and it offers an example of how a general system theory can be applied to all sorts of site phenomena. For the modern western system, Li selected Kevin Lynch’s system of site analysis, still the most comprehensive and widely accepted (by professionals) system of site analysis and design. By comparing these two systems a “universal” core of site systems was identified. From this core, a support system was developed as a general guide for planning, arranging the site, and subsequently improving and even conserving the land. The study is expected to open new possibilities for understanding designers, contractors, engineers, and other participants, rather than leaving them with specific site stereotypes. The project was undertaken in collaboration with Prof. P. Schmid and Prof. K. Ruitenbeek.

Automation Based Creative Design
Architectural theorists, educators, and practicing architects have often pointed out the importance of using precedent design products in the design process. But it is only recently that the role that precedents can play as design knowledge for design thinking and inference has begun to receive widespread attention by design computation researchers working to improve design practice. How far and in what manner precedents can be integrated in a computational design practice is a highly debatable subject. Most computer-based architectural techniques tend to suppress the importance of precedent. In collaboration with Dr. Ian White, Prof. Alexander Tzonis coordinated a research project and edited a book, presenting an up-to-date, comprehensive picture of research advances in the fast-growing field of Informatics applied to conceptual stages in the generation of buildings. The book addresses the question how far and in what ways creative design can be intelligently automated.
Twenty-four original contributions, introduced by a long critical essay, provide a balanced, yet challenging, forum of discussion. Among the topics covered are the use of precedents; the relations between case-based, rule-based, and principle-based architectural design reasoning; product typology; artifact thesauruses; the imputing and retrieval of architectural knowledge; the visual representation and recognition of built forms; empirical and analytical models of the design process and the design product; desktop design toolkits; grammars of shape and of function; design as a multi-agent collaborative process; grammars of shape and function; multi-perspective building data structures; design as a multi-agent collaborative process; and the integration of heterogeneous engineering information.

A Knowledge-Based Computational Approach to Architectural Precedent Analysis
Nan Fang in his Architectural Precedent Analysis (1993) developed a computational methodological frame work for the use of precedents. The research drew from the analysis of a case study of urban renewal in the ancient residential quarter of Beijing. The renewal project was carried out by the prominent Chinese architect and planner Prof. Liang Yong Wu, whose work relied heavily on the use of traditional precedents and who was also an advisor in this research. The case study was used heuristically to provide criteria for the devel opment of representations system of architectural precedents to be used in a design support system for advising architects in problems of integration of new modern structures in dense historical urban tissues. The new system relied on recent developments in cognitive theory, spatial semantics, and artificial intelligence. Among current methodologies in computer science, machine learning was used as a best-match tool to help reconstruct the design thinking implicit in the case study and, finally, to shape the new system.

C.J. Baljon (1993) developed a method to analyze conceptual systems contained in architectural discourses. He used texts by Ruskin, Semper, and Viollet-le-Duc as cases. The project was undertaken in collaboration with Prof. David van Zanten.

A. Zandi-Nia (1992) has developed TOPGENE, a computer-based system for generating preparametric spatial arrangements of buildings. The system employs deep models of expertise which can generate, evaluate, analyze, and diagnose malfunction in multiple social performance norms. The system uses Qanalysis to organize circulation operation data, and an iterative bottom-up approach in conjunction with hillclimbing and heuristic techniques, to arrive at a design. The research included analysis of the computational complexity of architectural design, the role of domain heuristics reducing such complexity and means for diagnosing the structural complexity of systems having a topological property represented as a graph, as is the case of designing buildings in relation to social performance norms. A new theorem for keeping track of distances in incrementally growing graphs was developed. The study also produced topological indices useful for quick diagnosis of alternative design solutions.
TOPGENE can generate various design types depending on the programmatic requirements. It uses rules of thumb to relate and code relations between form, operation and social performance norms in buildings combined with numerical data structures and computations. The system uses hill climbing to generate designs from a single point of view in conjunction with negotiation based problem solving. It employs an agenda mechanism to automate resolution of conflicts between the various programmatic social norms. Several experiments were carried out to test the system using realistic data. The results demonstrated the efficiency and the effectiveness of TOPGENE. In collaboration with Dr. E.J.H. Kerckhoffs he also compared the system with a neural network model borrowed from the Hopfield model. The superiority of TOPGENE was shown. TOPGENE has limitations as a totally automated system; parts of its methodology and special techniques can already be easily integrated in practice, in particular in fast pre-parametric evaluation of alternative building types. H.S. Roozenbeek contributed to the visualization of the of Zandi-Nia’s system.

The identification of the conceptual system contained in the early Arabic discourse on architecture was the topic of investigation of Badi Al-Abed (1992). The research also involved an extensive documentation of Arabic texts on architecture from the early Jahiliyya (pre-Islamic) period to the 18th century when foreign influences begin to be evident in the Arabic culture. The study discussed al-Jahiliyya verse, and religious, literary, scientific, geographical and historical texts, involving issues of building typology, urban planning principles, and descriptions of construction techniques, buildings and the belief systems with which they were built. Prof. I.R.T.M. Peters was advisor in the study.

Under a contract from the Rijksgebouwendienst of the Ministry of Housing and Planning J.S. de Boer and V. Mitossi prepared a typological study of morphological typological aspects of buildings related to performance criteria of flexibility and multi-functionality (1989).

A. Koutamanis (1990) has developed a framework for automated recognition of metric properties of architectural plans. Recognition includes identification of spatial primitives, grouping of primitives / subdivision of the plan and investigation of well-formedness. Image understanding and well-formedness concepts employed derived from the classical canon domain study researched by Dr. Liane Lefaivre and Prof. Alexander Tzonis (1986). The research is seen as a step towards the development of the module for inputting architectural precedent solutions in the Design Thesaurus, an automated intelligent database of precedents.

Methods and techniques for the management of the urban environment by means of image-processing using the computer and CAD systems were the research topic of Ronald Stenvert. The project included a systematic comparison of the prescriptions of the Classical orders given by twenty one treatises and a group of late sixteenth and seventeenth century Dutch buildings. The project was undertaken mainly at the University of Utrecht with Prof. Dr. J. van der Berg.

Garland Architectural Archives
To assist research in the generic process of design the Architectural Archives project was undertaken by Prof. Alexander Tzonis. Tzonis served as general editor of the Archives, which was published by Garland Publishing of New York. The Archives present the totality (not just a selection) of the corpus of documents in a known corpus of architect’s documents, identified and accompanied by standard assisting information . Among the architects whose archives were published were Le Corbusier, Mies van der Rohe (Award winning), Buckminster Fuller, Louis Kahn, Holabird & Roche and Holabird & Root (Award winning), Walter Gropius, F.L. Wrigh Index (Award winning), R.M.Schindler, Henri Sauvage and Alvar Aalto.